Continual misidentification of 5-oxolinalool as 2-methylnon-2-en-4-one in essential oils

ABSTRACT When employing GC and GC-MS, constituents of essential oils are most frequently identified by comparing retention indices and mass spectra with those reported in the literature or analytical libraries. While this allows for a rapid analysis, incorrect data can lead to a constituent being repeatedly misidentified. This was demonstrated in this work for 5-oxolinalool (6-hydroxy-2,6-dimethylocta-2,7-dien-4-one), an oxygenated monoterpene that was often erroneously identified as 2-methylnon-2-en-4-one. After encountering an unknown constituent of the essential oil of Sambucus nigra L. (Viburnaceae) flowers, we sought to acquire relevant analytical data on both compounds. Even though their mass spectra were rather similar (with 2-methylnon-2-en-4-one missing the peak at m/z 71), their retention indices were found to be significantly different. It follows that a revision of all prior identifications of 2-methylnon-2-en-4-one as an essential-oil constituent is needed.


Introduction
The developments in GC and GC-MS achieved in the 1980s have marked a turning point in the analysis of essential oils, a discipline now heavily reliant on hyphenated instrumental techniques and automated procedures (1).Identifications of known compounds are routinely made on the basis of retention indices and mass spectra matching against those of authentic samples (2).The confidence of a particular identification can be enhanced by acquiring and comparing RI data on two or more columns of varying polarity (3).Ultimately, a coinjection of the essential oil and an authentic sample of the analyte can offer proof of definitive identification and is welcome whenever such a sample is available (4).
While misidentifications in attempts to analyze essential oils might seem trivial, they can have far-reaching consequences, as the identifications of certain compounds in plant-derived materials could lead to their labeling as natural or nature-identical when used as ingredients in commercial products (5).Various misuses of GC and GC-MS data have also been documented, such as unreliable identifications of compounds that do not exhibit sufficiently specific mass spectra (e.g.sesquiterpenes) (6,7).Ailments include using additional spectroscopic techniques (IR and NMR) and strictly employing credible libraries compiled from physicochemical data on genuine samples, obtained either by isolation or synthesis (6,7).
Herein we present the case of the ongoing misidentification of 5-oxolinalool (6-hydroxy-2,6-dimethylocta-2,7-dien-4-one), an oxygenated monoterpene related to linalool, being systematically misidentified as 2-methylnon-2-en-4-one.As part of our studies on elderflower (Sambucus nigra L., Viburnaceae) volatiles (8,9), we identified by MS and RI matching what seemed to be 2-methylnon-2-en-4-one as a minor component of the oil of dried flowers.A closer look revealed that the structure of this compound is inconsistent with its previously reported RIs, and by determining the correct RI of this compound, as well as that of 5-oxolinalool, a compound exhibiting a similar mass spectrum, we demonstrated that a single wrong identification of this compound led to at least 16 erroneous reports of its existence in various essential oils.

General experimental procedures
Reagents and other chemicals were of analytical grade, commercially available (Sigma-Aldrich, St Louis, MO, USA, and Merck, Darmstadt, Germany), and were used as received unless stated otherwise.Hydrocarbon mixtures utilized for the determination of retention indices (RI) were purchased from Sigma-Aldrich (St. Louis, MO, USA).All solvents were HPLC grade and purchased from Sigma-Aldrich.Deuterochloroform (CDCl 3 ) was also acquired from Sigma-Aldrich.The 1 H and 13 C{ 1 H} NMR spectra, DEPT90, DEPT135, and 2D (NOESY, and gradient 1 H-1 H COSY, HSQC and HMBC) NMR spectra were recorded on a Bruker Avance III 400 MHz NMR spectrometer (Fällanden, Switzerland; 1 H at 400 MHz, 13 C at 101 MHz) equipped with a 5-11 mm dual 13 C/ 1 H probe head.All NMR spectra were measured at 25°C in CDCl 3 with tetramethylsilane (TMS) as the internal standard.Chemical shifts are reported in ppm (δ) and referenced to TMS (δ H = 0 ppm) in 1 H NMR spectra and to the solvent signal (δ C = 77.16ppm) in 13 C and heteronuclear 2D spectra.Built-in Bruker pulse sequences were employed.In the case of complex signals (overlapped and/or higher order), δ H and J values (expressed in Hertz) were manually adjusted to fit the experimentally available values and further optimized by spin simulation (10) in MestReNova (Mestrelab Research S.L., Spain; Figure S4).

Plant material and essential oil isolation
Umbels of S. nigra were collected from wild-growing populations in the vicinity of Šajkaš, Serbia, in full bloom on May 13, 2020.Inflorescences were not stripped of stems and were air-dried in thin layers on frames to constant mass.A voucher specimen was deposited in the Herbarium of the Faculty of Sciences and Mathematics, University of Niš, Serbia, under the acquisition number 17,753.The identity of the plant material was confirmed by the herbarium custodian.Dried elderflowers (500 g) were submitted to hydrodistillation for the duration of 2 h on a Clevenger-type apparatus.The obtained essential oil (ca.0.2 mL) was recovered by diethyl ether, which was subsequently dried with anhydrous magnesium sulfate and evaporated under a stream of nitrogen at ambient temperature and pressure.The essential oil yield was 180 mg (0.04%).

Gas chromatography analyses
The GC-MS analyses were carried out using a Hewlett-Packard 6890N gas chromatograph equipped with a fused silica capillary column (DB-5 MS; 5% diphenylsiloxane and 95% dimethylsiloxane, 30 m × 0.25 mm, film thickness 0.25 μm; Agilent Technologies, Palo Alto, CA, USA) and coupled with a 5975B mass selective detector from the same company.The analyses were performed in triplicate.The injector and interface were operated at 250°C and 300°C, respectively.The oven temperature was raised from 70°C to 290°C at a heating rate of 5°C min −1 , and the program ended with an isothermal period of 10 min.Helium was used as a carrier gas at a flow rate of 1.0 mL min −1 .The samples were injected as ethereal solutions (ca. 1 mg per 1 mL) in a pulsed split mode with a 40:1 split ratio.The flow was 1.5 mL min −1 for the first 0.5 min and then set to 1.0 mL min −1 throughout the remainder of the analysis.The mass spectrometer operated at an ionization voltage of 70 eV, in a 35-650 m/z acquisition mass range, with a scan time of 0.32 s.

Component identification
Aldol addition/condensation products were identified by comparison of their linear retention indices (RI) (11) relative to the homologous series of n-alkanes on a DB-5 MS column with literature data and their mass spectra with those of authentic standards, as well as those from NIST17 ( 12) and a homemade MS library with spectra corresponding to pure substances.The presence of 5oxolinalool was confirmed by a co-injection experiment with an authentic sample obtained by synthesis.

The crossed aldol reaction of acetone with heptan-2-one or 5-methylhexan-2-one
To a solution of heptan-2-one or 5-methylhexan-2-one (200 mg, 1.75 mmol) in acetone (10 mL) was added a 10% solution of potassium hydroxide in methanol (0.5 mL).After stirring overnight at room temperature, the reaction mixture was concentrated in vacuo to a volume of ca. 1 mL, which, upon dilution, was submitted to gas-chromatographic analysis.

Results and discussion
A gas-chromatographic analysis of the essential oil of dry elderflower leaves revealed it to be rather complex, with (Z)-hex-3-en-1-ol, hotrienol, linalool, and higher n-alkanes as prominent components; a detailed analysis of this volatile oil will be given elsewhere (9).A minor component eluting at an RI of 1215 (DB-5 MS column) caught our interest, as its mass spectrum was fairly similar to that of 2-methylnon-2-en-4-one, which a NIST library search returns as a match (Figure 1); the library entry is credited to Carl Djerassi.Indeed, this compound was reported with an almost identical retention index on a column of the same polarity, as even the NIST library entry states (14).
Our experience, however, suggested that the retention index was somewhat high for a simple ten-carbon enone.Since this compound is not commercially available, we opted for generating it in a non-preparative crossed aldol reaction of acetone and 2-heptanone so as to rapidly make a quick preliminary assessment about the correctness of our hypothesis on its retention index by running a GC-MS analysis of the reaction mixture.Expectedly, the crossed aldol reaction resulted in a multitude of products originating from self-and cross-additions/condensations.While products arising from acetone self-addition/condensation products dominated the reaction mixture, two chromatographic peaks corresponding to the acetone -heptan-2-one condensation and addition products were clearly recognizable.This simple experiment revealed the RI of 2-methylnon-2en-4-one (1180) to be too low to consider its identification at an RI of 1215.The corresponding aldol addition product featuring a tertiary alcohol group was observed at an RI of 1196, and no peak was evident at RI = 1215.Structural isomers of 2-methylnon-2-en-4-one are expected to display even lower retention on nonpolar columns; indeed, this was demonstrated by analyzing the product mixture obtained in a similar experiment where acetone was reacted with 5-methylhept-2-one, and the resulting 2,7-dimethyloctan-4-one was found to elute at RI = 1146 and displayed MS fragmentation patterns which were practically indistinguishable from those of 2-methylnon-2-en-4-one.
This finding implied that a new structure proposal was needed for the compound in question.The essential oil of Canella winterana (L.) Gaertn.leaves was reported to contain 2,6-dimethyloct-2-en-4-one (with no RI supplied) ( 15), another alkyl-chain isomer of 2-methylnon-2-en-4-one, but further considerations in this direction seemed futile on the basis of the RI values established on this occasion.The fact that aldol addition products have higher RIs relative to condensation products lead us to consider the possibility that another oxygenation is required to achieve the unknown constituent's RI.A biosynthetically reasonable assumption would be that it is most probably a structurally similar oxygenated monoterpene.The unknown compound's mass spectrum possessed a peak at m/z 71, frequently observed in mass spectra of monoterpenes such as linalool, corresponding to [C 4 H 7 O] + , the tertiary alcohol fragment ion.After considering several possibilities, we deemed the structure of the 5-oxo derivative of linalool (6-hydroxy-2,6-dimethylocta-2,7-dien-4-one) to be most likely.Such a structure would indeed account for the peak at m/z 71, which is prominent in the mass spectrum of the unknown compound, but barely observable in the spectrum of the reference substance (Figure 2).Additionally, peaks at m/z 83, 55, and 98 would also be expected.
This linalool derivative was first reported as a synthetic intermediate by Park et al. in 1977 (16), who prepared it by reacting the kinetic enolate of mesityl oxide generated by LDA at −78°C with methyl vinyl ketone.At a similar time, it was found as a volatile of Citrus junos Siebold ex Tanaka by Kitahara et al. (13); this work also includes details on the mass spectrum of 5-oxolinalool that is somewhat dissimilar from the minor constituent of S. nigra oil we encountered with respect to peak abundances but also features the m/z 71 ion.A subsequent study of the volatiles of S. nigra by a vacuum headspace technique revealed 5-oxolinalool to be one of the constituents (17).Having now a strong belief that the compound in question truly is 5-oxolinalool, we set out to seek definitive proof by synthesizing an authentic sample, following the procedure by Kitahara et al (13).
From the obtained sample it was easily determined that the mass spectrum of the compound was an excellent match (Figure S3), featuring all expected fragmentations, and the measured retention index was found to be 1215.A co-injection experiment also demonstrated the presence of 5-oxolinalool in the essential oil of dry S. nigra flowers.Considering this, as well as the information gathered on 2-methylnon-2-en-4-one, it is safe to conclude that the previous identifications of 2-methylnon-2-en-4-one on RI values of 1210-1220 are misguided and that, in all of these cases, the constituent in question is actually 5-oxolinalool.All literature identifications are tabulated below (Table 1); while the first known identification of the enone does not report a RI, the identification was based on MS spectra correlation with a prior version of the NIST library (18).It is noteworthy that the low abundance of 5-oxolinalool in all of the analyzed essential oils, including that of S. nigra, can be rationalized by its β-ketol structure, from which elimination of water can yield a highly conjugated ocimenone system, observed in S. nigra essential oil (16).Another possibility is that this structure degrades in a retroaldol process.

Conclusion
It was shown that the terpene alcohol 5-oxolinalool has been systematically misidentified as 2-methylnon-2-en-4-one in various sources, including at least 16 essential oils.The mass spectra of the two compounds are fairly similar, but not identical; 5-oxolinalool fragmentation leads to a peak at m/z 71, while no such peak is observable in the mass spectrum of 2-methylnon-2-en-4-one.However, the source of the confusion was traced to a single erroneous identification of 2-methylnon-2-en-4one at RI = 1215, a value later incorporated into NIST libraries.Here we demonstrated that the RIs of the two compounds are sufficiently different, with 2-methylnon-2-en-4-one eluting 35 units earlier.